17mm & 20/26/27mm single-sided DD/FET driver release: A17DD-SO8 / A20DD-SO8 / etc

Flashlight Modding and DIY Parts
1

Watch out for Zener build compatitibility issues in this version. There is a simple workaround for the current version (v024), see post #319 here for details. Also in post #319 and 320 you’ll find two different revised WIP board layouts which should both remove the need for any workarounds.

Here is a high power single sided FET driver. From my limited testing, fresh battery performance is similar to a regular BLF17DD. The driver is Zener ready. Note that this driver is not intended for a DPAK FET such as the (unavailable) Vishay 70N02; instead it uses an LFPAK56 (AKA Power-SO8) type FET like PSMN3R0-30YLD.

The primary goal for this driver was to create something which fit better in flashlights with retaining rings. The Supfire L6 is an example, it’s huge retaining ring prevents the use of two-sided driver. To that end this driver has absolutely no components on the spring side.

Secondary goals:

  • 1mm component keepout around the edge
  • Easy to build: reflow or hand solder easily (large 0805 sized pads)
  • Zener ready …sort of. See note at top of OP.
  • Offtime ready
  • Put my mark on it :wink:

More info:

  • All the above notwithstanding, this driver is still simply a derivative of the 17DD+7135 [WIP] driver I’ve been working on over here. The layout of the entire top half of the driver is the same.
  • As of v024 I used a 1.5mm thick GND ring on top and bottom where possible via polygons. Cereal_killer’s post #44 in the DD+7135 thread got me thinking and this is where I’m at for now. There is still only 0.5mm exposed. I feel like we’ve got plenty of strength now! Good thinking C_k.
  • D1/R3 is only designated through a triangle and a white line marking the cathode. The triangle has been updated slightly since v020 because it was too large and not sharp. The cathode mark was not present in v020 but should be present in v024.
  • Like the DD+7135 driver, this driver does have a reverse-polarity protection diode but bypasses the diode for the voltage divider. Due to this we are able to use any diode desired. Ideally we use a protection diode with an extremely low Vf in order to maximize gate voltage!
  • The exposed spring pad is 8mm. Scraping will reveal up to 11mm of copper. GND does not start until 15.5mm.

I’ve now built two of these using this component list:
C1 - 4.7uF [GRM21BR61E475KA12L]
R1 - 22k 0805
R2 - 4.7k 0805
D1 - SOD-323 & SOD-723, assorted
OTC - 1uf [C2012X7R1E105K125AB]
MCU - ATTINY13A-SSU
FET - PSMN3R0-30YLD

I have not assembled one with a Zener, but the pads are intended for an SOD-123 sized component. You may wish to use a higher capacity C1 than what I used, 10uF is fine as well. I used what I had on hand at the time. Based on some recent testing I did 3uF definitely does not work, 5uF and 4.7uF do work. I never tested 4uF. My 22k R1 replaces the 19.1k R1 due to the change in how the divider hooks up to B+. 22k is not guaranteed to be perfect, but it’s close enough for me and for now. Other more prolific modders may identify a closer value which appeals better. I’m happy with the cutoff I got from those components and a stock STAR offtime v1.3 firmware.

In testing with an OG XM-L of unknown bin I repeatably achieved 9A+ at the tailcap with a freshly charged Samsung INR18650-15M. The firmware uses JonnyC’s turbo rampdown mod, so current quickly falls of course. Here’s the dropin guts for that:

Finally, here are the OSH Park images.

17mm version:

https://oshpark.com/shared_projects/TrEN1wLB

20mm version:
w/ 1mm exposed copper on GND rings.

https://oshpark.com/shared_projects/YeNzvLPx

Credits to those who got the ball rolling on this stuff such as Mattaus and comfychair. And to the East-092 of course. :wink:

2

Reserved.

3

Very nice driver, thanks for the effort, I want those! I will have to find that FET first.

Also, and this is of course very important ;-) , this driver is extremely goodlooking with that FET placed in an angle and the triangular +pad in the middle :-)

4

+ 1 it’s the triangle that did it for me too.

I can’t wait to use these.

5

Oh wow…the DIE size of that MOSFET is almost the same size as the ATTiny13…daggum

So small yet so powerful

AWESOME!

What a great build! Just keeps getting better and better…Nice!

6

nice build!

7

Sweet! Thumbs up for single sided.

8

I've had luck taking the 3 amc nanjg variety drivers and removing the amc chips and then solder on some ??? ( I forgot what the package size that the amc chips are) and using that.

Just had to use 2 of them to achieve 12 A because the mosfet I used was a 6 amp max thingy.

I'll be looking to acquire some of these in the near future.. Would you be able to easily make a 20mm version too? (For Jacob A60 and I suppose others ;-) )

9

Thanks for the compliments folks! …I agree, the triangle is clearly the source of this driver’s Qi. :wink:

@ WarHawk-AVG - the FET’s plastic package (tisk tisk) is actually smaller than the ATtiny13A’s package. It’s not as wide or as long. When you include the length of the metal tab and pins I think the FET ends up being approximately 0.5mm longer, that’s it! They definitely fit on each other’s pads.

10

Yes, I can easily increase the size to whatever is appropriate. 20mm, anything else? I’ll take a look soon. I also took an unsuccessful stab at 15mm following a pre-release request. I was unable to get all the components to fit on one side in 15mm as I recall, so I moved the 3 lowest profile components to the bottom here.

11

DOH…right right :wink:

12

Very, very nice work! So we can use the parts list from the DD17 or should we use your list of parts? From mouser?

Really classy with everything on one side.

13

You may use the same parts you’ve been using with two exceptions:

  1. Most important: a DPAK sized FET will not fit. You must use LFPAK56/Power-S08, such as PSMN3R0-30YLD - Mouser / Digikey / Farnell (for the rest of the world).
  2. If you choose to use 19.2k for R1 instead of 22k you will need to adjust your LVP in the firmware before flashing. If you are using a pre-flashed ATtiny13A with LVP you must use 22k for R1. 19.1k will result in LVP allowing VERY low voltages. I highly recommend that you use 22k and be done with it. Digikey / Mouser / Farnell (for the rest of the world)

Some of our more determined friends here may prove that a DPAK FET can be made to fit upside down with some major filing. Basically the same filing you’d use to install it on a Nanjg-105c. IMO it won’t be worth the effort when you can install the DPAK FET on a Nanjg-105c or you can just drop in an LFPAK56 FET on my PCB no muss, no fuss.

If anyone wants to contribute any good eBay / Aliexpress links I’m all ears. :wink: Not everybody can buy through these distributors.

14

Here’s a link for 10 pcs.

Edit- couldn’t find data on this particular FET.

15

That is just awesome.

Really looking forward to ordering and trying to assemble some of these.

16

The problem with eBay links is it’s quite often hard to find the relevant data sheet for that particular FET and Rds on values can vary from absurdly low 1-2 mohms to more than 10mohms. Mouser at least gives data sheet links for each FET listed so you know what you are getting and will show the values for 4.5Vgs and 10Vgs as well as showing a graph from which you can interpolate behavior between 3-4V. Once again, the eBay crap shoot.

17

I think this is it: http://datasheet.elcodis.com/pdf/6/6/60632/ph4030al115.pdf

As you mentioned Vgs is important. In this case the Vgs graph just stops at 3.75v (where it’s still only <5milliohms but appears to be close to going vertical). Frankly that looks pretty similar to the FET I’m using as well as pretty similar to the DPAK AOD510.

Actually….

What about PSMN1R2-25YL? It’s late at night and I can’t go find the answer to how much gate charge the ATtiny13A can handle, but I think that that FET has a low enough gate charge (Qg) at <5v. The Rds(ON) figures look great to me, much better than the other FETs we’ve mentioned so far. I suspect that PSMN1R2-25YL would not be a good choice for Zener builds but might be very good for regular builds. I see them on eBay and I see that Mouser has them on order. Farnell has PSMN1R2-25YLC in stock, that’s not the same thing so watch out for that!

18

So, uh, here’s a thing I’ve never seen before…. http://cdn.vicorpower.com/documents/datasheets/Picor/ds_PI5101.pdf
It’s not actually smaller than LFPAK56, but the specs are crazy. Also they cost $5/ea, heh. And this is the only part I see on Digikey with that form factor.

19

Saw this design the other day when poking around oshpark wight… Didn’t know it was yours- I was thinking that’s kind of neat one sided board

20

Sweet, wight. Thanks.